Convex Optimization for Auxiliary Energy Management in Conventional Vehicles

Abstract

This submission deals with energy management of automotive ancillary systems, and how to find suitable control actions that can improve fuel economy compared to control laws that do not exploit the storage potential of energy buffers. The main focus lies on how to model components to fit within the framework of convex optimization. The resulting convex models allow the energy-management problem to be formulated as a second-order cone program. Higher levels of a control hierarchy deliver an estimate of optimal costates to a lower level, where the problem is repeatedly solved with the recently developed Embedded Conic Solver (ECOS). The control law is verified and compared with a “standard” control law in simulations.